Hadamard integrator for time-dependent wave equations: Lagrangian formulation via ray tracing

08/18/2023
by   Yuxiao Wei, et al.
0

We propose a novel Hadamard integrator for the self-adjoint time-dependent wave equation in an inhomogeneous medium. First, we create a new asymptotic series based on the Gelfand-Shilov function, dubbed Hadamard's ansatz, to approximate the Green's function of the time-dependent wave equation. Second, incorporating the leading term of Hadamard's ansatz into the Kirchhoff-Huygens representation, we develop an original Hadamard integrator for the Cauchy problem of the time-dependent wave equation and derive the corresponding Lagrangian formulation in geodesic polar coordinates. Third, to construct the Hadamard integrator in the Lagrangian formulation efficiently, we use a short-time ray tracing method to obtain wavefront locations accurately, and we further develop fast algorithms to compute Chebyshev-polynomial based low-rank representations of both wavefront locations and variants of Hadamard coefficients. Fourth, equipped with these low-rank representations, we apply the Hadamard integrator to efficiently solve time-dependent wave equations with highly oscillatory initial conditions, where the time step size is independent of the initial conditions. By judiciously choosing the medium-dependent time step, our new Hadamard integrator can propagate wave field beyond caustics implicitly and advance spatially overturning waves in time naturally. Moreover, since the integrator is independent of initial conditions, the Hadamard integrator can be applied to many different initial conditions once it is constructed. Both two-dimensional and three-dimensional numerical examples illustrate the accuracy and performance of the proposed method.

READ FULL TEXT

page 28

page 29

page 30

page 33

page 34

page 35

page 36

page 37

research
08/14/2021

Global space-time Trefftz DG schemes for the time-dependent linear nonhomogeneous and anisotropic wave equation

In this paper we are concerned with Trefftz discretizations of the time-...
research
05/25/2023

Local Randomized Neural Networks with Discontinuous Galerkin Methods for Diffusive-Viscous Wave Equation

The diffusive-viscous wave equation is an advancement in wave equation t...
research
01/14/2021

A Machine-Learning Method for Time-Dependent Wave Equations over Unbounded Domains

Time-dependent wave equations represent an important class of partial di...
research
12/18/2020

Data-driven rogue waves and parameter discovery in the defocusing NLS equation with a potential using the PINN deep learning

The physics-informed neural networks (PINNs) can be used to deep learn t...
research
10/18/2022

Spectral solver for Cauchy problems in polar coordinates using discrete Hankel transforms

We introduce a Fourier-Bessel-based spectral solver for Cauchy problems ...
research
12/21/2019

Shadow Lagrangian dynamics for superfluidity

Motivated by a similar approach for Born-Oppenheimer molecular dynamics,...
research
03/04/2021

Time-dependent stochastic basis adaptation for uncertainty quantification

We extend stochastic basis adaptation and spatial domain decomposition m...

Please sign up or login with your details

Forgot password? Click here to reset